Automated Reconstruction, Evaluation and Comparison of Diverse Genome Scale Fungal Models

Authors: 
Edirisinghe, J. N., Argonne National Laboratory
Faria, J. P., Argonne National Laboratory
Liu, F., Argonne National Laboratory
Calhoun, S., Lawrence Berkeley National Laboratory
Grigoriev, I., Lawrence Berkeley National Laboratory
Henry, C. S., Argonne National Laboratory
Genome-scale metabolic models offer an efficient mechanism for predicting phenotypes across many environmental conditions. These models are also key tools in understanding interactions between species in the environment, including microbial, plant, and fungal species. While methods have emerged for the automated reconstruction of microbial and plant models from genomic data, the construction of high-quality fungal models remains a challenge. Here we introduce a new method for constructing genome-scale fungal models in an automated fashion based on a curated set of reactions and gene associations that are derived from 14 published fungal metabolic models. As the basis for our method, we produced a fungal model template that encompasses the biochemistry data from the published fungal models and the structural annotations from the associated fungal genomes. Using structural annotations of a fungal genome we compute a set of orthologous proteins against the curated fungal template in order to assert the presence or absence of specific biochemical reactions and pathways. Once the orthologous protein families are determined, the related biochemistry data is propagated to construct a new draft metabolic model. This method is deployed in the KBase (https://narrative.kbase.us/) as an app called "Build Fungal Model". We have built tools in KBase to produce high-quality metabolic models on prokaryotes and plants and new approaches to identify trophic interactions between species. Using these methods, we are able to study how fungal-bacterial, plant-fungal and microbiomes evolve as a sustainable community.

We applied “Build Fungal Model” app to build draft models for more than 100 fungal genomes imported from the Joint Genome Institute (JGI) MycoCosm resource (https://genome.jgi.doe.gov/programs/fungi/index.jsf).

We compared these models side-by-side, exploring how each genome overlaps with our curated model template and plotting model variance along the phylogenetic tree of fungal genomes. All draft fungal models are available for download via a KBase Narrative.